CONNECTING LINK AND SEMICONDUCTOR DEVICE

Abstract

A connecting member (1) connects a first target object (2) and a second target object (3). The connecting member (1) includes: a metal particle (11) containing Ni as a main component; a low-melting-point phase (12) containing Sn as a main component and having a melting point of less than 300°C; and an intermetallic compound (13) having a melting point of 300°C or more and formed by interdiffusion between Sn and the metal particle (11). A ratio of an amount of the low-melting-point phase (12) to a total amount of the connecting member (1) is 2% by volume or more and less than 20% by volume. Each of the thermal expansion coefficients of the first target object (2) and the second target object (3) is 3 × 10 -6 / K or more and less than 13 × 10 -6 / K, and a difference between the thermal expansion coefficient of the first target object (2) and the thermal expansion coefficient of the second target object (3) is less than 5 × 10 -6 / K. The thermal expansion coefficient of the connecting element (1) is 16 × 10 -6 / K or more and less than 20 × 10 -6 / K.

Description

TECHNICAL FIELD

[0001] The present disclosure relates to a connector and a semiconductor device. STATE OF THE ART

[0002] In recent years, there has been an increasing demand for improving the reliability of semiconductor devices. In particular, there has been a need to improve the reliability of the lifetime of a bonded portion between components that have different thermal expansion coefficients (for example, a bonded portion between a semiconductor element and a circuit board, or a bonded portion between a circuit board and a heat dissipation plate).

[0003] Conventionally, semiconductor devices, each comprising a substrate using silicon (Si), gallium arsenide (GaAs), or the like, are widely used. The operating temperature of each of these semiconductor devices is 100°C to 125°C. A solder material used for bonding between the semiconductor device and the printed circuit board must have a high melting point to withstand multi-step solder bonding during manufacturing; crack resistance to repeated thermal stress caused by starting and stopping; and resistance to device contamination.

[0004] In response to these requirements, for example, 95Pb-5Sn (mass%) was used as the solder material for Si devices and 80Au-20Sn (mass%) was used as the solder material for gallium arsenide devices.

[0005] However, 95Pb-5Sn, which contains a large amount of harmful lead (Pb), poses a problem in terms of reducing its environmental impact. On the other hand, 80Au-20Sn, which contains a large amount of precious metal, poses a problem in terms of its increased price or the limited amount of precious metal. Therefore, alternative materials for both are highly desired.

[0006] On the other hand, with a view to energy savings, a device using a substrate made of silicon carbide (SiC) or gallium nitride (GaN) has been actively developed as a next-generation device. The operating temperature of each of these is 175°C or higher, and is expected to reach 300°C in the future.

[0007] This means that the temperature of the bonded portion between the semiconductor element and the circuit board reaches 175°C. Furthermore, the temperature of the bonded portion between the circuit board and the heat dissipation plate also increases to a temperature close to this, depending on the operating condition and heat dissipation performance. Therefore, high reliability is required for the bonded portion between the circuit board and the heat dissipation plate, as well as the bonded portion between the semiconductor element and the circuit board.

[0008] As described above, in each of the conventional devices and next-generation devices, a member (connecting member) having a high melting point and excellent heat resistance has been required to form each of the connected portions.

[0009] As a conventional method for manufacturing a joint body having high heat resistance, the following method is known: a large amount of metal particles such as Ag or Cu is added to Sn, and a joint layer is formed by an Ag-Sn alloy phase (for example, Ag3Sn: melting point of 480 °C), a Cu-Sn alloy (for example, Cu6Sn5: melting point of 415 °C; Cu3Sn: melting point of 676 °C) or the like by metal diffusion caused by heating (300 °C or less) at the time of joining.

[0010] Regarding an assembly method or an electronic device using solder, for example, PTL 1 (JP 2002-314241) discloses a connection structure in which Sn of a solder ball is melted to form an intermetallic compound at an interface with Cu of metal balls, and the Cu metal balls are coupled to each other. Note that since the molten Sn also forms intermetallic compounds with an electrode of a semiconductor chip, an electrode of an intermediate substrate, and the like, the Cu metal balls and these electrodes are coupled.

[0011] Furthermore, PTL 1 describes that a joining portion can ensure sufficient strength to withstand a process at the time of subsequent joining by soldering, even if Sn of the solder ball becomes a Cu-Sn intermetallic compound (Cu6Sn5; melting point: about 630 °C) to increase the melting points of a contact portion and its surroundings, and a part of Sn remains unless the other parts are melted.

[0012] It should be noted that PTL 1 also describes that: the metal balls are not limited to Cu, respectively; Ag, Au, Al, Ni, a Cu alloy, a Cu-Sn compound, an Ag-Sn compound, an Au-Sn compound, an Al-Ag compound, a Zn-Al compound, or the like can be used therefor; and Au has excellent wettability and therefore has a void reducing effect at the joining portion.

[0013] Furthermore, PTL 2 (WO 2012 / 108395) describes that a solder containing an Sn-based metal and a Cu-based metal is heated and melted at a joining portion, thereby forming an intermetallic compound having a melting point of 310°C or more. Note that it is also disclosed that a ratio of the Sn-based metal component contained in the joining portion is 30% by volume or less. LIST OF CITIONSPATENT LITERATURE PTL 1: JP 2002 - 314 241 PTL 2: WO 2012 / 108395 SUMMARY OF THE INVENTION TECHNICAL PROBLEM

[0014] However, in the conventional method for manufacturing a semiconductor device as illustrated in each of PTL 1 and PTL 2, an amount of the metal particles added to the Sn particles is not evaluated with respect to crack resistance in a thermal shock test.

[0015] For example, when a large amount of the metal particles such as Cu particles or Ag particles is added, the fluidity of Sn deteriorates and a portion sealed by the metal particles therefore becomes a non-joined portion, generating a void, with the result that a crack is likely to be generated when a thermal shock is applied to the joined portion.

[0016] Furthermore, in terms of a metal phase diagram, the Ag-Sn alloy represented by Ag3Sn or the Cu-Sn alloy represented by Cu6Sn5 has about 5 mass% of a solid solution region of Ag or Cu and Sn. However, since there is a separation state of two phases, an interface between Ag3Sn and Cu6Sn5 becomes a void unless there is a Sn phase with a low melting point, and becomes the starting point of a crack.

[0017] On the other hand, the inventors of the present invention considered that in a semiconductor device operating at a high temperature, it is important to ensure not only a mechanical property as a material or heat cycle resistance as the only component applied thereto, but also reliability as a device in a power cycle (repeating generation of heat and cooling by repeating turning on and off the application of power to a semiconductor device) that an actual operation includes.

[0018] An object of the present disclosure is to provide an interconnection member having high interconnection reliability for a power cycle in a semiconductor device operating at a high temperature. SOLUTION TO THE PROBLEM

[0019] A link connects a first target object and a second target object.

[0020] The connecting member includes: a metal particle containing Ni as the main component; a low melting point phase containing Sn as the main component and having a melting point of less than 300 °C; and an intermetallic compound having a melting point of 300 °C or more and formed by interdiffusion between Sn and the metal particle.

[0021] A ratio of an amount of the low melting point phase to a total amount of the linking member is 2 volume% or more and less than 20 volume%.

[0022] Each of the thermal expansion coefficients of the first target object and the second target object is 3 × 10 -6 / K or more and less than 13 × 10 -6 / K and a difference between the thermal expansion coefficient of the first target object and the thermal expansion coefficient of the second target object is less than 5 × 10 -6 / K.

[0023] The thermal expansion coefficient of the connecting link is 16 × 10 -6 / K or more and less than 20 × 10 -6 / K. ADVANTAGEOUS EFFECTS OF THE INVENTION

[0024] According to the present disclosure, it is possible to provide a connector having high connection reliability for a power cycle in a semiconductor device operating at a high temperature.

[0025] Specifically, in the connecting member of the present disclosure, the low melting point phase (for example, a phase including only Sn) remains to ensure wettability at a connecting target member (connecting target object), and a remaining amount thereof is optimized to suppress a vertical crack in the power cycle.

[0026] Further, since the connecting member of the present disclosure is applied to connect the plurality of connection target objects between which the difference in thermal expansion coefficient falls within the specific range, it is possible to suppress a lateral crack caused by a shear stress generated by the difference between the thermal expansion coefficients of the plurality of connection target objects.

[0027] Furthermore, since the connecting member having a thermal expansion coefficient falling within the specific range is used, tensile stress and compressive stress (tensile-compressive stress) caused by the difference between the thermal expansion coefficient of the joining target and the thermal expansion coefficient of the connecting member are suppressed, thereby achieving excellent connection reliability even when operating at a high temperature (for example, 175 °C). BRIEF DESCRIPTION OF THE DRAWINGS Fig. 1 is a flowchart showing a process for manufacturing a semiconductor device (an interconnection member). Fig. Figure 2 is a schematic cross-sectional view showing a first step of Fig. 1 shows. Fig. Figure 3 is a schematic cross-sectional view showing a second step of Fig. 1 shows. Fig. Figure 4 is a schematic cross-sectional view showing a third step of Fig. 1 shows. Fig. Figure 5 is a schematic cross-sectional view showing a fourth step of Fig. 1 shows. Fig. Figure 6 is a schematic cross-sectional view showing a fifth step of Fig. 1 shows. DESCRIPTION OF EMBODIMENTS

[0028] Embodiments of the present disclosure are described below. It should be noted that in the figures, dimensional relationships such as length, width, thickness, and depth have been appropriately changed for clarity and simplicity of the figures and do not represent actual dimensional relationships. First embodiment. <verbindungsglied>

[0029] A connecting member 1 of the present embodiment is a member that connects a first target object 2 (for example, a heat radiation plate) and a second target object 3 (for example, a circuit board having a semiconductor element and a wiring circuit) (see Fig. 6). It should be noted that the connecting member 1 is different from a solder foil 10 described later and is a member made of solder (solder alloy) that is solidified after being melted by a heating step and that is in a state where the connecting member 1 is connected to both of the connection target elements (first target 2 and second target 3).

[0030] The connecting member 1 of the present embodiment includes metal particles 11, a low melting point phase 12, and an intermetallic compound 13.

[0031] Each of the metal particles 11 contains Ni as a main component. Here, the "main component" is a component that has the largest amount among components contained in the metal particle 11. The content ratio of Sn in the metal particle 11 is preferably 60 mass% or more, more preferably 70 mass% or more, and further preferably 80 mass% or more.

[0032] The metal particle 11 may further contain at least one component selected from a group consisting of Fe, Cr, C, Cu and Si.

[0033] The low melting point phase 12 contains Sn as a main component and has a melting point of less than 300°C. Here, the "main component" is a component that has the largest amount among the components contained in the low melting point phase 12. The content ratio of Sn in the low melting point phase 12 is preferably 60 mass% or more, more preferably 70 mass% or more, and further preferably 80 mass% or more.

[0034] The low-melting-point phase may further contain, for example, Ag and Cu (in addition to Sn). In this case, the content ratio of Ag may be, for example, 3 mass% or more and less than 4 mass%, and the content ratio of Cu may be 0.5 mass% or more and less than 1.0 mass%. Note that when the content ratios of Ag and Cu, respectively, fall within these ranges, the effect of the present disclosure is expected to be achieved with greater certainty.

[0035] The intermetallic compound 13 is a compound (phase) formed by interdiffusion between Sn and the metal particle 11 and has a melting point of 300 °C or more.

[0036] Further, in the connecting member 1 of the present embodiment, a ratio of an amount of a low melting point phase 12 to a total amount of the connecting member 1 is 2 volume % or more and less than 20 volume %.

[0037] Furthermore, each of the thermal expansion coefficients of the first target object 2 and the second target object is 33 × 10 -6 / K or more and less than 13 × 10 -6 / K, and a difference between the thermal expansion coefficient of the first target object 2 and the thermal expansion coefficient of the second target object 3 is less than 5 × 10 -6 / K.

[0038] Furthermore, the thermal expansion coefficient of the connecting link is 16 × 10 -6 / K or more and less than 20 × 10 -6 / K.

[0039] It is to be noted that, as a result of conducting a reliability test on a semiconductor device in a power cycle, the inventors of the present invention found that, in the power cycle, suppression of cracks (vertical cracks) generated in the layer of the interconnect member in a thickness direction is important for improving interconnect reliability.

[0040] Furthermore, the reliability of the connected portion depends not only on the material of the connecting link, such as solder, but also on the configuration of the connection target link (connection target). Therefore, to improve the reliability of a power module, it is considered necessary to check not only the composition of the connecting link material but also the configuration of the connection target link.

[0041] It should be noted that regarding vertical cracks, a reference document (Koji Yamazaki, "A Study on Vertical Crack Initiation of Lead-Free Solder Joints," 26th JIEP Annual Meeting, 9C-14) describes that when a high-speed thermal cycle test (+50 °C / +175 °C; temperature difference: 125 °C) was performed on a sample in which Cu plates were joined together by soldering and a sample in which Mo plates were joined together by soldering, no vertical crack was generated in the former sample and a vertical crack was generated in the latter sample, although the same solder material was used in both samples for the purpose of joining.

[0042] This is due to the tensile and compressive stresses (tensile-compressive stress) caused by the difference between the thermal expansion coefficient of the solder material and the thermal expansion coefficient of the joining target member. When restrained by a member with low thermal expansion, the tensile-compressive stress is applied to the solder, which has a high thermal expansion coefficient. Furthermore, dynamic recrystallization occurs in the solder due to heating, resulting in the generation of a vertical crack along a grain boundary of Sn, the main component of the solder, or along an internal precipitate.

[0043] It is noted that when the difference between the thermal expansion coefficients of the connection target members (two connection target objects) is large, it is known that mainly a lateral crack is generated instead of a vertical crack.

[0044] As a result of the above-described examination, the inventors of the present invention have found that the connector of the present disclosure has high connection reliability for a power cycle in a semiconductor device operating at a high temperature. Second embodiment. <halbleitervorrichtung>

[0045] A semiconductor device according to the present embodiment includes a heat-radiating plate 2 and a circuit board 3 (a substrate including a semiconductor element and a wiring circuit). The heat-radiating plate 2 and the circuit board 3 are connected with the connector 1, which has high connection reliability as described in the first embodiment, being interposed therebetween. Thus, a semiconductor device with high reliability is provided.

[0046] The semiconductor device of the present embodiment can be manufactured by a manufacturing method including a step of joining the heat radiation plate and the circuit board (see Fig. 1 to 6).

[0047] With reference to a Fig. 1, an exemplary process for connecting the circuit board and the heat radiation plate in the method for manufacturing a semiconductor device according to the present embodiment will be described below.

[0048] First, in a first step (S1), the solder foil 10, in which metal particles 11 consisting of Ni are dispersed in the low melting point phase 12 containing Sn as the main component and having a melting point of 300 °C or less, is cut into a predetermined size (see Fig. 2).

[0049] Specifically, as a representative example of the solder composition of the low melting point phase 12, a composition represented by Sn-3.3Ag-0.5Cu (the numerical values ​​each represent mass% and Sn is the remainder of the composition ratios of the other elements) was achieved by adding and dissolving each element, thereby preparing a base body.

[0050] Next, a solder foil 10 was prepared in the form of a roll by a rolling process at 100 °C or less so that it has a thickness of 100 µm.

[0051] Next, Ni particles (metal particles 11) were evenly dispersed in the solder foil 10, and a foil consisting of the low-melting-point phase 12 was overlapped on top of it, thereby producing a mixed foil. In the present embodiment, its thickness was 100 μm, but this is not limited to this. Furthermore, its cut size was 10 mm × 10 mm, which was the same as that of the upper member.

[0052] It should be noted that, in addition to rolling, the solder foil 10 can be manufactured by melting the base body and then introducing the molten solder through a thin slit. Furthermore, when more than three elements are combined as described above (for example, when Bi, In, Sb, or the like are added to Sn-Ag-Cu), the composition of the base body may become uneven due to temperature variation in a melting furnace. In this case, the solder composition can be adjusted by first preparing a base body of Sn-Ag-Cu, then redissolving it, and adding a predetermined amount of the remainder of Bi, In, Sb, or the like. Alternatively, the solder foil 10 can be manufactured by first forming solder balls, uniformly dispersing the solder balls on a flat plate, and performing compression molding.

[0053] Next, in the second step (S2), the solder foil 10 is placed on the first target object 2 (lower link) (see Fig. 3).

[0054] It should be noted that, to prevent oxidation, a predetermined amount of an anticorrosive agent having a decomposition temperature of 100°C or less and not affecting the bonding property may be applied to the outermost surface of the CuMo alloy plate. Furthermore, when placing the solder foil 10, an organic agent that thermally decomposes at a high temperature of 100°C or more may be used as an adhesive material to prevent the solder foil 10 from deviating from the predetermined position. The viscosity of the adhesive material is preferably 200 Pa s or more. Alternatively, a cover film may be applied to portions other than the bonded portion of the first target object 2 to prevent the solder foil 10 from deviating.

[0055] The solder foil 10 may be slightly curved at the time of cutting or handling. This is due to the following reason: Since the solder foil 10 is melted to some extent by heating, the influence of the initial shape of the solder foil is small. However, if the solder foil 10 is severely curved to the extent that the second target object 3 (upper member) cannot be placed thereon, the solder foil 10 is corrected on another flat plate so that it is parallel to it, since the second target object 3 (upper member) is to be placed thereon in the next step.

[0056] Next, as a third step (S3), the second target object 3 (upper link) is placed on the solder foil 10 (see Fig. 4).

[0057] It should be noted that the adhesive material described above may be applied between the connection target member (upper member) 1 and the solder foil 10 to fix the mounting position.

[0058] It should be noted that each of the surfaces of the connection target members (first target 2 and second target 3) on the connection side is preferably provided with Cu or Ni plating, or preferably coated with an Au film, an Ag film, a Pt film, or the like in the order of several tens of nm.

[0059] Next, as a fourth step (S4), the stack of the respective members obtained in the above-described steps was placed on a heating plate 5 in a heating furnace 5 (see Fig. 5). Then, formic acid, which is a representative organic acid capable of reducing an oxide film, was sealed in the heating furnace 5, and the solder foil 10 was melted by heating at 180°C for 5 minutes and then heating at 260°C for 3 minutes, thereby forming the connecting member 1.

[0060] Here, the low-melting-point phase 12 and metal particles 11 in the solder foil 10 form a compound due to thermal diffusion. The formation ratio of the compound (intermetallic compound) of the metal particles 11 and Sn varies depending on the particle sizes and the amount of addition of the metal particles 11. Therefore, it is preferable to adjust the particle sizes and the amount of addition of the metal particles 11 so that the low-melting-point phase 12 remains.

[0061] One reason for this is as follows: First, when a plurality of portions of the connecting member 1 are formed by metal particles 11 and the intermetallic compound 13, Sn of the low melting point phase 12 cannot be deformed, and therefore there is no problem in connection reliability.

[0062] On the other hand, each of the surfaces of universal joining target members (first target 2 and second target 3) may have roughness, and the surface roughness (unevenness) may be on the order of several μm. Therefore, when joining is performed without applying pressure, an amount of the low-melting point phase 12 corresponding to the volume of the unevenness of the surface must also wet and spread. However, when a large amount of metal particles 11 is present, the surface area of ​​the metal particles 11 in the solder foil 10 becomes large, the low-melting point phase 12 preferentially reacts with metal particles 11, and the low-melting point phase 12 does not sufficiently wet the surface of the joining target member.

[0063] Therefore, in the present embodiment, it is preferable to adjust the particle sizes and the amount of addition of the metal particles 11 so that the low-melting-point phase 12 remains. For example, the particle sizes (average particle size: D50) of the metal particles 11 added to the solder foil 10 are each 5 to 20 μm, and a mixing ratio of the metal particles in the solder foil 10 is 8 to 30 mass%.

[0064] It should be noted that if there is no concern about damage to the connection target links, the connection can be performed by applying pressure.

[0065] Next, as a fifth step (S5), the sample having the formed connecting member 1 was placed on a cooling plate 6 for the purpose of cooling (see Fig. 6).

[0066] On this occasion, in a case where the solder connecting member is made of a material that is completely melted, such as ordinary solder, there is a possibility that a connection target member (the upper member or the lower member) is disadvantageously warped or curled (unevenness is formed) due to thermal contraction. Therefore, it is desirable to perform cooling in a stepwise manner in the cooling step. For example, cooling to 100°C is generally performed within 60 seconds; however, to reduce warpage, cooling to 100°C may be performed over about 400 seconds.

[0067] Since the main component of the metal particles 11 in the present disclosure is Ni, which has a thermal expansion coefficient lower than that of Sn, an amount of thermal contraction of the solder foil 10 is small, the thermal expansion coefficient of each of the connection target members (first target 2 and second target 3) is 3 × 10 -6 / K or more and less than 13 × 10 -6 / K and the difference between the thermal expansion coefficient of the first target object 2 and the thermal expansion coefficient of the second target object 3 is less than 5 × 10 -6 / K, with the result that the distortion can be small and the residual stress generated in the connecting link 1 can also be small. Therefore, the cooling step time does not need to be particularly long.

[0068] In an actual application of this to a power module, for example, a printed circuit board including a semiconductor element and a circuit pattern formed thereon may be bonded to a heat radiation plate with the connecting member 1 interposed therebetween. The connecting member 1 (solder foil 10) may be applied to a die pad portion, a base mounting portion, or the like of the power module regardless of the size of the bonding area.

[0069] It should be noted that the oxygen concentration in the solder foil 10 is also important. Each of the metal particles 11 contains Ni, which is likely to be oxidized, and the surface area of ​​the metal particles 11 is large. Therefore, even if the areas of a small amount of the metal particles 11 are oxidized, oxidation proceeds from the interfaces of the metal particles 11, resulting in a large amount of oxidation overall, thereby greatly affecting wettability. Therefore, if the oxygen concentration in the solder foil 10 is high, an excellent bonding member cannot be formed unless it is stored in a vacuum desiccator or the like. Therefore, to form an excellent bonding member, the oxygen concentration in the solder foil 10 is preferably more than 0 and less than 500 ppm. Examples

[0070] Hereinafter, the present disclosure will be described in more detail with reference to examples, but the present disclosure is not limited to these examples. (Test example 1)

[0071] Compound samples (samples 1 to 7) were tested with the method described in the above-described embodiments with reference to Fig. 1, wherein the thermal expansion coefficients of the first target objects 2 and second target objects 3 were changed as shown in Table 1. It should be noted that the particle size (average particle size: D50) of the metal particles 11 added to each solder foil 10 is 10 µm.

[0072] A CuMo alloy plate (thickness: 1 mm; size: 20 mm × 20 mm) was used as each first target 2. The outermost surface of the CuMo alloy plate was pure copper and was not coated with any plating. In these CuMo alloy foils, the mixing ratios of Cu and Mo were changed to change the thermal expansion coefficients, as shown in Table 1. Note that an Invar alloy, which has a thermal expansion coefficient of 2, was used as a first target.

[0073] It should be noted that in the case of a DBC (Direct Bonded Copper) or a CIC (Copper-Invar-Copper) substrate, the thermal expansion coefficient can be appropriately adjusted by the thicknesses of the three layers included in the substrate.

[0074] Each second target 3 has a thickness of 100 µm and a size of 10 mm × 10 mm. The thermal expansion coefficients of the second targets 3 were changed as shown in Table 1 according to the mixing ratios of the CuMo alloy and the Invar alloy in the same manner as for the first targets 2.

[0075] In order to prevent the wettability from differing depending on the type of the connection target member, Ni plating having a thickness of 3 µm was applied to each of the surfaces of the first target 2 and the second target 3 on the connection side.

[0076] The joint samples of the combinations shown in Table 1 and obtained as described above were subjected to a high-speed thermal shock test simulating a power cycle test. Specifically, the thermal shock test was performed by repeatedly performing 100,000 cycles, with each cycle consisting of 10 seconds of electrical conduction (reached temperature: 175 °C) and 10 seconds of electrical conduction (reached temperature: 50 °C).

[0077] After the thermal shock test, a cross-section of the link layer of each joint sample was formed in the thickness direction, and a transmission X-ray image of the cross-section was binarized for image analysis, whereby a ratio of the area of ​​the voids generated in the link to the area (100%) of the entire link in the cross-section was measured. If the ratio of the void area was 10% or more, it was evaluated as "NOK," whereas if the ratio of the void area was less than 10%, it was evaluated as "OK" because there was no practical problem. The evaluation results are shown in Table 1.

[0076] [Table 1] Coefficient of thermal expansion [10 -6 / K] Bewertungsergebnis Second target object (3 in Fig. 6) First target object (2 in Fig. 6) Differenz Probe 1 2,0 6,0 4,0 NIO Probe 2 3,0 7,3 4,3 OK Probe 3 7,3 11,8 4,5 OK Probe 4 12,5 17,3 4,8 OK Probe 5 12,5 18,4 5,9 NIO Probe 6 7,3 12,5 5,2 NIO Probe 7 11,8 17,2 5,4 NIO

[0078] Considering the results shown in Table 1, it is understandable that when each of the thermal expansion coefficients of the first target object 2 and the second target object 3 is 3 × 10 -6 / K (3.0 × 10 -6 / K) or more and less than 3 × 10 -6 / K (13.0 × 10 -6 / K) and the difference between the thermal expansion coefficient of the first target object 2 and the thermal expansion coefficient of the second target object 3 is less than 5 × 10 -6 / K (5.0 × 10 -6 / K), the evaluation result is OK, and voids are less likely to be generated in the connector after the high-speed thermal shock test. Therefore, according to the present disclosure, it is understood that it is possible to provide a connector having high connection reliability for a power cycle in a semiconductor device operating at a high temperature; and a semiconductor device using the connector.

[0079] It should be noted that these results are also related to the thermal expansion coefficient of the connected portion (the connecting member and the connection target member). As described above, a tensile-compressive stress in the lateral direction is applied to a connecting member restrained by connection target members (first target 2 and second target 3) to generate a crack in the vertical direction. That is, when the thermal expansion coefficient of the connecting member 1 is small, the generated stress may also be small. In the present disclosure, metal particles 11 each containing Ni as a main component are used. It is difficult to predict from public documents to what extent the thermal expansion coefficient of the connecting member 1 employing metal particles 11 affects connection reliability.

[0080] As a result of the actual verification of the thermal expansion coefficient and the like of the connecting member 1 by the inventors of the present invention, it was found that excellent results (samples 2 to 4 in Table 1) can be obtained when the thermal expansion coefficient of the connecting member is 16 × 10 -6 / K or more and less than 20 × 10 -6 / K and the ratio of the area of ​​the low melting point phase (phase containing Sn as the main component and having a melting point of less than 300 °C) to the area (100%) of the entire link in the cross section of the link (cross section of the layer of the link in the thickness direction) is 2% or more and less than 20% (that is, when the ratio of the amount of the low melting point phase 12 to the total amount of the link 1 is 2 vol% or more and less than 20 vol%).

[0081] The embodiments and examples disclosed in this specification are illustrative and not restrictive in all respects. The scope of the present disclosure is defined by the content of the claims, not by the embodiments described above, and is intended to include any modifications that fall within the scope of the claims and the meaning equivalent to their content. LIST OF REFERENCE SYMBOLS

[0082] 1 connecting link; 10 solder foil; 11 metal particles; 12 low melting point phase; 13 intermetallic compound; 2 first target object (heat radiation plate); 3 second target object (circuit board); 4 heating furnace; 5 heating plate; 6 cooling plate. QUOTES CONTAINED IN THE DESCRIPTION

[0000] This list of documents submitted by the applicant was generated automatically and is included solely for the convenience of the reader. This list is not part of the German patent or utility model application. The DPMA assumes no liability for any errors or omissions. Cited patent literature

[0000] JP 2002 - 314 241 [0010, 0013] WO 2012 / 108395

[0013] Cited non-patent literature

[0000] Koji Yamazaki, “A Study on Vertical Crack Initiation of Lead-Free Solder Joints,” 26th JIEP Annual Meeting, 9C-14

[0041] < / halbleitervorrichtung> < / verbindungsglied>

Claims

[1] A connecting member for connecting a first target object and a second target object, the connecting member comprising: a metal particle containing Ni as the main component; a low melting point phase containing Sn as the main component and having a melting point of less than 300 °C; and an intermetallic compound having a melting point of 300 °C or more and formed by interdiffusion between Sn and the metal particle, where a ratio of an amount of the low melting point phase to a total amount of the linking member is 2% by volume or more and less than 20% by volume, each of the thermal expansion coefficients of the first target object and the second target object 3 × 10 -6 / K or more and less than 13 × 10 -6 / K and a difference between the thermal expansion coefficient of the first target object and the thermal expansion coefficient of the second target object is less than 5 × 10 -6 / K and a thermal expansion coefficient of the connecting link 16 × 10 -6 / K or more and less than 20 × 10 -6 / K. [2] The connecting member according to claim 1, wherein the low melting point phase further contains Ag and Cu. [3] A connecting member according to claim 1 or 2, wherein the metal particle further contains at least one component selected from a group consisting of Fe, Cr, C, Cu and Si. [4] Connecting member according to one of claims 1 to 3, wherein the first target object is a heat radiation plate, and the second target object is a printed circuit board having a semiconductor element and a wiring circuit. [5] A semiconductor device comprising: a heat radiation plate; and a circuit board having a semiconductor element and a wiring circuit, wherein the heat radiation plate and the circuit board are connected with the connecting member according to claim 4, which is arranged between the heat radiation plate and the circuit board.

Images